Impact of pregravid obesity on maternal and fetal immunity: fertile grounds for reprogramming.

Abstract

Pregravid obesity results in several complications during pregnancy and is a major determinant of fetal outcomes and offspring health. Specifically, pregravid maternal obesity and in utero exposure to high fat diet have been shown to have detrimental effects on fetal programming, predisposing the offspring to adverse cardiometabolic, endocrine, and neurodevelopmental outcomes. More recently, a deeper appreciation for the modulation of offspring immunity and infectious disease-related outcomes by maternal pregravid obesity has emerged. This review will describe currently available animal models for studying the impact of maternal pregravid obesity on fetal immunity and review the data from clinical and animal model studies. We also examine the burden of pregravid obesity on the maternal-fetal interface and the link between placental and systemic inflammation. Finally, we discuss future studies needed to identify key mechanistic underpinnings that link maternal inflammatory changes and fetal cellular reprogramming events.

Introduction

Almost 36% of women of childbearing age in the US are categorized as obese (BMI > 30) [1]. Several studies have linked high maternal BMI before/during pregnancy with a higher incidence of several obstetric and post-partum complications, including gestational diabetes [2, 3], gestational hypertension, preeclampsia (PE) [4, 5], higher need for caesarean delivery [6], microbial infections (urinary and genital tract infection, and sepsis) [7–11], chorioamnionitis [12, 13], and post-partum surgical site infections [14–17]. Pregravid (pre-pregnancy) obesity also poses significant risks to the fetus including early pregnancy loss, preterm delivery [18, 19], stillbirth [20], and delivery of large-for-gestational-age infants [21].

In line with the developmental origins of health and disease (DOHaD) hypothesis, epidemiologic studies have linked pregravid obesity with detrimental cardiometabolic, neurocognitive, and behavioral outcomes in the offspring [22–28]. More recent studies have also revealed a significant impact on the fetal immune system indicated by higher incidence of bacterial and viral infections in neonates born to mothers with obesity that require admission to the neonatal intensive care unit [29–31]. Moreover, the higher incidence of pregravid obesity has been implicated in the rising incidence of allergic diseases (childhood wheeze and atopy) [32] and asthma [33–35] during early life [36, 37], childhood [38–40], and adolescence [41]. We aim to provide an overview of the impact of pregravid obesity on 1) maternal and offspring immunity, 2) the maternal-fetal interface, and 3) cellular and molecular mechanisms that mediate these changes.

Modeling maternal obesity.

Ex vivo studies using human samples provide the ideal system to study the impact of maternal obesity on fetal immunity [42, 43]. In clinical studies, obesity is defined as a body mass index (BMI) that exceeds 30kg/m². Clinical studies have the advantage of collecting longitudinal data on fat mass, gestational weight gain, and post-partum weight retention/loss as well as identifying the areas of fat deposition that provide a comprehensive picture of the obesogenic perinatal environment [44]. However, studies involving human samples can also be challenging to control and interpret. Firstly, it is not clear if the experimental readouts in human studies are direct consequences of obesity or mediated by comorbidities associated with obesity. Although large epidemiological studies are successful in establishing associations between maternal BMI and adverse offspring outcomes, they do not provide precise mechanisms of action. Finally, it is not feasible to study the impact of maternal obesity on immune cells in maternal/fetal tissue compartments. Nevertheless, the placenta, cord blood, and the microbiota provide practical avenues to study fetal reprogramming.

The use of animal models obviates some aforementioned hurdles and allows a more mechanistic approach to identify molecular underpinnings of observed outcomes [45]. One major challenge in animal models is establishing the duration of nutritional insult, which is critical in mimicking outcomes observed in human subjects. In this section, we briefly describe the more frequently used animal models for studying the impact of maternal obesity on both the maternal and fetal immune systems. Extensive details of these models have been reviewed in [45–47]. Rodent models are often used to study developmental reprogramming, owing to short gestation, genetic homogeneity, ease of genetic manipulation, and access to multiple tissues [48]. These studies have linked maternal overnutrition using high fat (HFD) or western diet (WD) to poor fetal growth, metabolic syndrome, and behavioral outcomes in the offspring [49–51]. Rat models using Sprague-Dawley and Wistar strains have also been used to study the effects of nutritional status during pregnancy on short term and long-term offspring health [52–54]. In contrast to mouse strains, rats are outbred thereby providing greater genetic heterogeneity and offer the advantage of having larger placentas and fetuses. Finally, rodent models have been developed to study the impact of maternal adiposity on fetal health, independent of dietary fat intake [55]. However, several key differences exist between rodent and human immune systems. First, unlike humans, splenic hematopoiesis persists in rodents for several weeks after birth [56]. Secondly, T and B cell development is not complete at birth and several subsets of T lymphocytes are absent in 10 days old pups, suggesting an immature adaptive immune system at birth [57].

These limitations can be in part overcome by the use of larger animal models. For example, sheep are precocial species like humans and have been used to study fetal growth and metabolic consequences in the context of maternal obesity [58, 59]. Furthermore, sheep models of maternal pregravid obesity have demonstrated increased offspring predisposition to cardiac [60] and hepatic dysfunction [61]. However, sheep have a distinct placentation compared to humans and hence, may not be ideal for studying immune development at the maternal-fetal interface.

Non-human primates (NHPs) are closely related to humans and have been used extensively to study fetal-placental adaptations. NHPs are born with a fully functional immune system that shares significant homology with that of humans, making them a particularly attractive alternative for in utero immune reprogramming studies. Research groups have leveraged baboons [62], rhesus macaques [63–65], and Japanese macaques [66] to model adiposity and inflammation both on the maternal and fetal fronts. Although rhesus macaques have a bidiscoid placenta [67], baboons have very similar placental structure as humans [68].

Early immune development: a window of opportunity for in utero reprogramming

The profile of the immune system and the magnitude of its responses vary significantly with gestational and post birth age. Here, we provide a brief overview of the ontogeny of the fetal immune system, highlighting striking differences in its responses from that of adults (Figure 1). Greater details can be obtained from these additional reviews [69–71]. Fetal hematopoiesis occurs in distinct spatial and temporal sites: the extraembryonic yolk sac, the fetal liver, thymus, and bone marrow. Primitive erythroid macrophage, and megakaryocyte lineage cells appear in blood islands of the yolk sac as early as day 18 of gestation [72]. By gestational week 3–4, hematopoietic cells appear in the fetal liver, reaching roughly 60% of the liver cells between weeks 7 and 15. While neutrophils are present in the human fetal liver parenchyma as early as gestational week 5 [73], monocytes appear in circulation by week 6 following seeding of newly formed self-renewing HSCs in the fetal liver [73, 74]. Monocyte progenitors then colonize various organs and compartmentalize into specialized tissue-resident macrophages that can persist for life [73]. The frequencies of monocytes and dendritic cells (DCs) increase over gestation. However, neutrophils are the most prominent immune cell subtype at birth.

Figure 1:

Overview of the ontogeny of the human immune system and composition of circulating immune cells at birth.

Fetal lymphoid organs (spleen, thymus, and lymph nodes) appear between 8–12 weeks. Mature B-lymphocytes can be detected in the fetal liver at 8 weeks of gestation, then in circulation by week 12 [75], and the primary follicles of the lymph nodes at week 17 [76]. Peripheral and marginal zone B cells then undergo somatic hypermutation forming a diverse B cell receptor repertoire [77]. By contrast, T cell progenitors begin migrating to the thymus at gestational week 8–9 and circulating mature fetal α T lymphocytes are detected at 14 weeks of gestation, several months after detection of subpopulations of innate lymphoid cell subsets. The fetal T cell repertoire continues to diversify significantly during the second and third trimester [77]. Finally, by week 20, human HSCs migrate from the liver to the bone marrow, which takes over as the primary production site for red blood cells, lymphoid progenitors, mature B cells, and myeloid cells. Absolute numbers of major lymphoid subsets stabilize at gestational week 20–26 and are comparable to the numbers seen in newborns [78].

Neonatal environment instructs distinct immune functions

Both intrinsic and extrinsic cellular factors instruct the neonatal immune response to be poised towards a predominantly Th2 phenotype. Cord blood and neonatal serum levels of anti-inflammatory cytokines (IL-4, IL-10, IL-13, and TGF-b) are higher than those observed in adults, supporting a Th2 biased response [79]. Additionally, cord blood contains a higher concentration of immunosuppressive adenosine, which inhibits neutrophil activation [80] and TLR mediated production of Th1 cytokines [81]. However, fetal skin and the mucosal layers produce high levels of antimicrobial proteins and peptides (APPs) such as defensins, lactoferrin, lysozyme, and cathelicidins, which limit infection [82–84]. Moreover, serum levels of hepcidin, which reduce free iron, are high at birth thereby limiting bacterial growth and reducing the risk for neonatal sepsis [85, 86].

Innate Immunity

At birth, neutrophils exhibit quantitative and qualitative differences compared to their adult counterparts [87]. Cord blood and infant neutrophils are poorly phagocytic [88], less chemotactic [89], produce less APP and reactive oxygen species (ROS) in response to pathogenic encounters [90, 91], and exhibit impaired ability to form neutrophil extracellular traps (NETs) compared to their adult counterparts [92]. Monocytes in newborns are primarily classical (CD16^low) and produce lower levels of pro-inflammatory cytokines (TNF-α) and Th1 promoting cytokines (IL-12, IFNα), higher levels of IL-10 and IL-1RA [93], but comparable or higher levels of Th-17 promoting cytokines (IL-6, IL-23) compared to adult monocytes [94].

Newborns have fewer conventional DCs (myeloid; mDCs) at birth [95], which progressively increase over the first five years of life [96]. Like monocytes, they are less responsive to TLR stimulation than adult DCs, with significant impairment in IL-12 production [97, 98]. Furthermore, quantitative differences in surface expression of costimulatory molecules CD80 and HLA-DR also exist [99]. Neonatal mDCs also are highly susceptible to external factors such as regulatory B cell suppression [100]. On the other hand, numbers and responses of circulating plasmacytoid dendritic cells (pDCs) are comparable in newborns and adults [101]; however, both neonatal mDCs and pDCs demonstrate reduced ability to generate a polyfunctional response to TLR stimulation. [102].

These age-associated differences in myeloid cell responses have been attributed to cell-intrinsic factors such as: (a) high levels of intracellular cAMP in neonatal cells, which suppress Th1 but enhance Th2 cytokine production [103]; (b) changes in chromatin accessibility that regulate binding of transcription factors such as IRF3 and subsequent cytokine expression [104]; and (c) alterations in histone modifications (trimethylation on H3K4) within promoters of inflammatory genes marking activation [105]. Human cord blood is also highly enriched with immunosuppressive CD71+ erythroid cells [106] that regulate T cell activation and cytokine release from myeloid cells [107]. Frequencies of these cells correlate with gestational age and with gestational age-dependent susceptibility to neonatal infections [108]. Finally, newborn blood is also enriched with myeloid-derived suppressor cells (MDSCs) that play a critical role in controlling inflammation.

Effector functions of innate lymphocytes such as invariant Natural Killer T (iNKT) cells, mucosal-associated invariant T (MAIT) cells [109], and gamma delta (γδ) T cells [110] appear to be lower in early life. Natural Killer (NK) cell responses such as degranulation and release of lytic factors are reduced early in neonatal/infant cells compared to adults, explaining increased infant susceptibility to viral infections [111]. This age-associated immaturity in NK cell and innate lymphoid cells has been attributed to TGF-β mediated suppression [112].

Adaptive immunity

The adaptive immune system of the newborn is fully developed but lacks antigenic experience. Specifically, newborns and infants have fewer effector memory T cells (CD45RO+CD45RA-) and B cells (CD27+) [113, 114]. These populations increase with age and antigenic experience. In a newborn, the majority of αβ T lymphocytes are recent thymic emigrants (RTE) [115] and exhibit a distinct epigenetic program compared to adult mature T cells [116, 117]. More often, newborn T cells are epigenetically poised to mount Th2-dominant responses (IL-4 and IL-13) [118–120] and the cytokine profile of neonatal APCs (IL-6 and IL-23) support their robust ability to mount Th-17 responses [121]. Over time, the newborn’s immune system undergoes a gradual shift from Th2 to Th1 responses characterized by the production of IL-12p70 by DCs, promoting cellular immunity and macrophage activation [69].

Neonatal T cells can generate limited Th1 responses characterized by lower IFNγ production by T cells and Th1 polarizing cytokines by APCs [122]. The magnitude of Th1 responses is determined by the quality of signals present at the time of T cell priming and antigen presentation [123]. Similarly, multiple pathogens including CMV, HIV, and T. cruzi can induce effector CD8 T lymphocytes in the neonate [124, 125], albeit with reduced cytotoxicity, and reduced production of antimicrobial peptides and ROS [126]. Additionally, fetal T cells are prone to a more rapid onset of functional exhaustion [127]. Finally, newborns and infants have higher frequencies of regulatory T cells [128] but with reduced capacity to suppress DC function [129]. As described for T cells, neonatal B cells generate reduced primary antibody responses to vaccines and infections due to dampened BCR signaling, decreased plasma cell differentiation and germinal center responses, and lower DC activation signals [130].

Evidence of fetal immune reprogramming by pregravid obesity

Recent studies have documented that the adverse impact of maternal pregravid obesity on fetal cellular developmental processes extend beyond the cardiovascular and central nervous system and have a significant impact on phenotypic and functional aspects of offspring immunity (Figure 2). Clinical studies are limited to the profiling of cord blood, which serves as an excellent surrogate for a snapshot of the fetal immune system at birth [98]. Animal models, on the other hand, facilitate long-term studies and the study of various tissue-resident cells.

Figure 2:

Pleiotropic impact of high pregravid BMI on circulating and tissue resident immune cells as well as lung, adipose tissue, liver, gut, and brain.

Impact of Maternal obesity on phenotypic and functional changes in cord blood cells

There is increasing evidence from human studies linking pregravid maternal BMI and changes in immune cell frequency and phenotype in cord blood [42, 131, 132]. One recent study reported a drop in eosinophils and CD4+ T cells in the cord blood [42] with pregravid obesity after controlling for several maternal factors (race, maternal infection, gestational diabetes, gestational hypertension, and preeclampsia). A second retrospective study reported that pregravid maternal BMI positively correlated with frequencies of NK T cells and regulatory CD8+ T cells and negatively correlated with that of CD34+ stem cells in cord blood samples [131]. More recent work reported higher numbers of immunosuppressive nucleated RBCs [132] in cord blood of babies born to mothers with obesity.

From a functional standpoint, maternal BMI is negatively correlated with cord blood CD4+ T cell IL-4 production [42]. Additionally, cord blood monocytes are hypo-responsive to ex vivo stimulation with TLR agonists in the absence of any differences in the relative frequencies of classical and non-classical monocytes [42] or differences in surface expression of CD14, CD16 or TLR4 [43]. This phenotype was consistent when purified monocytes were used, suggesting a cell-intrinsic defect [43]. A more recent study reported that cord blood monocyte-derived macrophages (MODM) from babies born to mothers with obesity showed a basal anti-inflammatory phenotype and exhibited unbalanced responses to M1 and M2 polarization stimuli [133]. These in vitro findings are line with recent clinical data indicating increased incidences of severe bacterial infections [134] and sepsis [135] necessitating NICU admission in neonates born to mothers with BMI >30 [29, 136] and highlight the importance of cord blood monocytes as critical early pathogen sensors and phagocytes in the neonate. Surprisingly, pregravid maternal BMI did not impact resting gene expression in cord blood monocytes, suggesting a significant role for epigenetic poising. Indeed, dampened monocyte responses to LPS were mirrored at the transcriptional level and was partially explained by loss of DNA cytosine methylation at promoters of inflammatory genes such as PPARG, FOLR2, and ITGAX [43].

Poor ex vivo immune responses in the offspring could potentially be tied to an altered inflammatory intrauterine environment. Very high maternal BMI (>35) has been associated with elevated systemic levels of inflammatory mediators, notably TNFα and CRP in the cord plasma [137, 138]. In addition to these canonical inflammatory markers, maternal BMI is also a strong indicator of cord blood levels of malondialdehyde (MDA) and nitric oxide (NO), contributing to increased oxidative stress in the newborn [139]. In agreement with this hypothesis, transcriptional studies in cord blood have demonstrated elevated expression of genes involved in the cellular response to oxidative stress and inflammatory signaling with high maternal BMI, independent of offspring adiposity [140]. Gene expression studies in NHP models have extended these findings and reported altered expression of several markers of vascular inflammation and altered endothelial cell function in offspring of dams with high pregravid BMI [141]. Additionally, PBMC isolated from offspring of obese and lean baboons (based on pre-pregnancy BMI) revealed widespread transcriptional changes in genes inherent to antigen presentation pathways, complement and coagulation pathways, leukocyte trans-endothelial migration, signaling pathways (BCR signaling, MAPK, and VEGF signaling) [62]. These observations provide evidence of fetal reprogramming events that persist during early childhood.

Maternal obesity and alterations in offspring lymphoid tissue resident cells

Studies using a variety of animal models have assessed the impact of pregravid maternal obesity on tissue-resident immune cells in the offspring (Figure 2). Some of the earliest studies using mice demonstrated that the pups born to dams fed a HFD during gestation had an altered immune system regardless of postnatal diet [142, 143]. One key study concluded that gestational HFD exposure results in fewer splenic lymphocytes and thinner thymic cortex [143]. More importantly, pups born to dams fed a HFD diet during gestation exhibit impaired antigen-specific immune responses as evidenced by: (a) higher mortality rates in an E. coli sepsis model; (b) increased bacterial burden; (c) larger abscesses following Methicillin-resistant Staphylococcus aureus (MRSA) infection [143]; and exacerbated respiratory disease following infection with respiratory syncytial virus (RSV) [144, 145]. Furthermore, in utero HFD exposure resulted in reduced production of IL-6 and TNFα by splenocytes, but exacerbated IL-17, IL-6, and TNFα production by colonic lamina propria lymphocytes (LPLs) [143]. Both in the spleen and the colon, in utero HFD exposure resulted in reduced frequencies of regulatory T cells [143]. In line with that observation, pups born to obese dams were also more likely to develop signs of experimental autoimmune encephalitis (EAE) [143]. Finally, offspring with HFD exposure both during gestation and post weaning produced higher antigen-specific (ovalbumin) IgE and lower IgG following in vivo immunization, potentially linking maternal pregravid obesity to increased risk of allergic responses and autoimmunity in the offspring [142].

Maternal obesity and liver, adipose tissue, and gut homeostasis in the offspring

Several lines of evidence suggest that maternal obesity significantly reprograms the liver of the offspring, contributing to metabolic disease. Exposure to an obesogenic diet both during gestation and post weaning contributes to higher frequencies of Kupffer cells in the offspring liver. Furthermore, Kupffer cells in the offspring of HFD-fed dams were less phagocytic and produced higher levels of ROS [146]. Gene expression studies in infant rhesus macaques exposed to HFD during gestation provide substantial evidence of oxidative stress in fetal livers during the third trimester as evidenced by elevated expression of metabolic enzymes and lipotoxicity [63]. Furthermore, cellular changes in fetal livers are observed as early as 8–12 weeks of gestation, with increased recruitment of pro-inflammatory macrophages in fetuses of dams exposed to WSD [147]. Similarly, in utero exposure to WSD in mice resulted in reprogramming of bone marrow derived macrophages to be hyper-responsive to LPS, as evidenced by elevated Il1b, Tnf, Il6, and Il10 transcript levels [147].

More recent work has mechanistically linked liver inflammation in the offspring of mothers with obesity to alterations in the gut microbiome. Specifically, mice that received a fecal transplant from two-week old human babies born to mothers with obesity exhibit histological evidence of peri-portal inflammation, impaired macrophage phagocytosis, and dampened cytokine production, all hallmarks of non-alcoholic fatty liver disease (NAFLD) [148]. Furthermore, hepatic gene expression signatures in recipient mice suggest ER stress and reprogramming of innate immune cells [148]. Similar phenotypes have also been reported in adipose tissues, where offspring of obese dams had higher CD68, CCR2, and TNFA transcript levels, indicative of a shift towards a more inflammatory phenotype potentially due to increased infiltration of macrophages [149]. Finally, reduced expression of Glut4 gene in the sub-cutaneous fat of these offspring suggests potential metabolic reprogramming [149].

In addition to the liver and adipose depots, exposure to HFD during gestation alters inflammatory responses in the fetal gut. Specifically, rodent studies have reported enhanced susceptibility to DSS induced colitis in the offspring that was accompanied by up regulation of NF-κB signaling and expression of Il1b, Il6, and Il17 genes [150]. This phenotype was associated with increased neutrophil infiltration in the gut, paralleled with higher CCL2 transcripts [150]. A sheep model of maternal obesity reported similar changes in the fetal gut as early as the second trimester, with increased expression of monocyte/macrophage genes TNF, IL1B, IL6, IL8, CCL2, CD68, ITGAM (encoding CD11b), and CD14 in fetal large intestine [151]. Furthermore, both mRNA and protein levels of TGF-β were elevated in gut tissue obtained from offspring of obese ewes. TGF-β plays a critical role in maintaining immune homeostasis and regulating interactions between the microbiota and lymphocytes in the gut [152]. Moreover, increased levels of TGF-β could potentially predispose the offspring to inflammatory conditions such as IBD [153]. In line with these observations, mice exposed to gut microbes from two-week-old infants born to mothers with obesity had impaired gut barrier indicated by reduced gene expression of tight junction proteins Tjp1 and Ocln and increased intestinal permeability [148]. In addition, infants born to mothers with obesity show dysbiosis in their gut microbiome [154, 155] for up to 2 years [156].

Finally, mouse models have recently revealed that the impact of maternal obesity extends to microglia in the offspring brain. Specifically, CD11b+ cells from offspring born to obese dams exhibit exaggerated TNFα responses to LPS [157]. This observation is in line with an increased incidence of autism, and Attention Deficit/Hyperactivity Disorder (ADHD) in offspring of mothers with obesity [158–161].

Pregravid obesity and changes in offspring stem cells

The overall fitness of hematopoietic stem cell (HSC) populations is extremely sensitive to an obesogenic environment. Using a combination of colony formation assays, flow cytometry, and gene expression experiments, studies showed that maternal HFD restricts physiological expansion of fetal HSCs, compromises repopulation of precursor cells in fetal liver, and biases cellular differentiation towards the myeloid branch [162]. This bias towards myeloid lineage and dysregulated hematopoiesis was shown to be dependent on toll-like receptor 4 (TLR4) signaling [163]. Interestingly, metabolic rewiring and increased energy demand of stem cells have been recently reported in mesenchymal stem cells isolated from human cord blood samples obtained from babies born to mothers with obesity [164].

Pregravid obesity and inflammation at the maternal-fetal interface

A successful pregnancy involves complex interactions between the immune cells residing at the maternal-fetal interface [165]. Structurally, the interface between the uterine mucosa and extraembryonic tissue comprise the “maternal-fetal interface” or the placenta. At term, this interface includes the maternal portion called the decidua (or decidua basalis), which is primarily the uterine mucosal layer after it has undergone implantation-associated differentiation process called decidualization. The fetal portion of the placenta is called the villous chorion, which allows for the transfer of nutrients from maternal blood to fetal circulation. During the first trimester, the placenta enables the establishment of pregnancy, providing physical support and immunologic tolerance to the developing embryo [166]. As gestation progresses, the placenta takes over the roles of facilitating nutrients, waste, and gas exchange, and produces hormones and other factors that support fetal growth. Finally, closer to parturition, the cells in the placenta promote contraction of the uterus, the expulsion of the baby and rejection of the placenta [167]. In this section, we briefly summarize the impact of pregravid obesity on placental structure and function as well as the role of these changes in reprogramming the fetal immune system (Figure 3).

Figure 3:

Impact of obesity on maternal systemic immunity and adaptations at the maternal-fetal interface.

Data from animal model studies show that consumption of HFD increased placental glucose, amino acid, and fatty acid nutrient transport in dams [53, 168–171]. Evidence for altered trophoblastic invasion and vessel remodeling have been reported in a rat model of HFD induced obesity [172, 173]. Furthermore, placentas of rhesus macaque dams fed a HFD exhibited reduced blood flow, increased calcification and risk for infarction [66]. A reduction in placental microvessel density has been observed as early as mid-gestation in a mouse model of diet-induced obesity [174]. In obese human subjects, significant structural changes in the placenta have been observed, ranging from alterations in placental maturity, vessel density, muscularity in blood vessels [175] and non-branching angiogenesis in term placentas [176].

Inflammation in the placenta has a bidirectional effect on both the maternal and fetal immune systems [177]. Transcriptional studies using human term placenta have reported changes in genes regulating inflammation, lipid storage and transport [178], insulin resistance, as well as angiogenesis in both the decidua [179] and villi [180, 181] with pregravid obesity. Similarly, RNA-Seq analysis of villous tissue revealed down-regulation of several genes involved in inflammation (IL1R2, IL2RB, C3, TNFSF10), immune tolerance (HLAG), cell matrix communication, nutrient transport, and retinoic acid metabolism with pregravid obesity [181]. While the source of placental inflammation is not clear, it has been linked to the high levels of oxidative stress and lipotoxicity in the decidua of mothers with obesity [180]. In the next section, we describe the consequences of pregravid obesity on the immune landscape at the maternal-fetal interface.

Obesity associated adaptations in the placenta: impact on the microbiome and immune cell landscape

A combination of 16S rDNA and whole-genome shotgun metagenomic analyses has showed that the placenta harbors a low abundance but metabolically rich microbiome [182] that varies from the decidua to basal plates to the fetal membranes [183]. This microbial community is disrupted with preterm labor [184], preeclampsia [185] and gestational weight gain [186], and pregravid obesity [181]. Some of these studies were carried out using placenta samples of preterm births; therefore, our understanding of the impact of pregravid obesity on microbial communities within the term placenta remains limited.

Technical advances in single cell RNA sequencing has allowed for high throughput profiling of placenta-resident immune cells in an unbiased manner in the early placental bed [166, 187]. These studies indicate the presence of NK cells, macrophages, and T-cells as well as differences in origin, phenotype, and function of immune cells within the decidua and villous compartments. For example, the decidual immune cells, which are of maternal origin, are primarily composed of NK cells, macrophages, and T cells. Immune cells in the villi, which are of fetal origin, are primarily composed of macrophages called “Hofbauer cells” [166]. Both diversity and phenotypes of placenta-resident immune cells have been shown to considerably shift over the course of gestation [188–190]. For instance, studies have shown a shift in placental cytokine milieu (decreased IL-35 and IL-9) and concurrent drop in CD11c+ myleloid cells and regulatory T cell populations over the course of gestation [188].

Local changes in frequencies, phenotype, and functions of immune cells in the decidua and villi have been attributed in the pathogenesis of several obstetric complications such as preterm labor [191], preeclampsia [192–194] and chorioamnionitis [195]. Given the increased frequencies of these pathologies with high pre-pregnancy BMI, it is widely believed that alterations in placental immune cells could be the drivers. Indeed, first trimester placentas from women with obesity show reduced uterine NK cell numbers, which normally comprise ~70% of the leukocyte population in early human decidua and play a central role in uterine artery remodeling and altered expression of genes in pathways associated with matrix remodeling and growth factor signaling [196]. Furthermore, first trimester uNK cells from women with obesity exhibited an imbalance in expression of NK cell receptors KIR2DL1 (inhibitory receptor) and KIR2DS1 (activating receptor), ultimately favoring HLA-C2 directed activation [197]. Finally, leptin levels have been linked to altered expression of CD56 and CD16 on circulating NK cells during pregnancy [198]. Therefore, it is possible that obesity-associated increase in leptin levels could also modulate phenotype and function of uterine NKs, however, this hypothesis has yet to be tested.

Macrophages comprise 20% of the decidual leukocyte population. These cells are distinct in their origin but are functionally similar to their fetal counterpart called the “Hofbauer cells”. Decidual macrophages play important roles in maintaining immune tolerance [199], protecting the fetus against infectious agents [200], spinal artery remodeling [201], and clearance of apoptotic bodies from the placental vasculature [202]. Although phenotypic and functional changes in the decidual macrophage populations over the course of gestation remain ill defined, these cells are extremely heterogeneous [203]. In humans, pregravid obesity results in an increased accumulation of macrophages (CD68+) in the placenta [175, 204]. In line with these observations, higher transcript levels of pro-inflammatory genes IL6, CCL2, IL8, and TLR4 are detected in a baboon model of HFD-induced obesity [66, 151, 205]. Similarly, chorionic villi macrophages obtained from placentas of diet-induced obese dams generate exacerbated responses to LPS (higher TNFα, IL-6, and IL-1β) both at the gene and protein levels, suggesting functional reprogramming of fetal macrophages as well [157]. A more recent study demonstrated reduced frequency of M1-like macrophages (HLA-DR+ CD163-) in the decidua with pregravid obesity [206], potentially a compensatory mechanism against obesity-associated inflammation.

Pregravid-obesity induced changes in systemic maternal immunity

Given that a number of immune cells in the placental bed (NK cells, T cells, and subsets of macrophages) are recruited from peripheral blood during gestation, BMI-associated alteration in inflammatory pathways in the placenta could potentially be linked to systemic changes. Indeed, several studies have linked alterations in frequencies and phenotypes of immune cells in the placenta with circulating cytokines and markers of inflammation. In this section, we briefly describe pregnancy-associated alterations in the maternal circulating immune system and disruptions triggered by an obesogenic environment.

Recent proteomic and immunological studies in peripheral blood have suggested the presence of an “immunological clock of pregnancy” characterized by enhanced immune activation with gestation [207, 208]. Changes in the maternal adaptive immune include decreased Th1/Th17 and increased Th2/Treg responses [207, 209, 210] as well as increased surface expression of CD11a, CD11b, CD49d, CD14, CD64, and CD54 on monocytes during third trimester [211–213]. Moreover, the production of pro-inflammatory cytokines IL-1β, IL-12, IL-6 and oxygen free radicals following stimulation with LPS or bacteria [207, 214, 215] as well as responses to viral particles by NK cells [216], monocytes [207] and plasmacytoid dendritic (pDC) cells [217] were enhanced with gestation. While the mechanisms regulating these changes are still unknown, several studies have hypothesized a role for placental microparticles [218], fetal DNA [219], and/or maternal metabolic and cytokine changes [208, 220]. However, it should be noted that these studies were largely carried out using peripheral blood mononuclear cells (PBMC) and therefore intrinsic changes within specific immune cell subsets remain largely ill defined.

Obesity exacerbates inflammatory changes normally associated with pregnancy [221] (Figure 3). While most studies have consistently reported increased IL-6 levels with higher pregravid BMI in pregnant women [9, 175, 222–229], others have reported no associations at all [230–232]. These inconsistencies arise due to sample size limitations, variability in exclusion criteria, racial differences in cohorts, and the precise time of sample collection (first, second, third trimester, at term, during labor or post-partum). On the other hand, all studies report a significant increase in CRP, a canonical marker of inflammation produced by the liver, during various stages of pregnancy with pregravid BMI [9, 181, 204, 223–225, 229, 232–234]. Pregravid obesity also impacts circulating levels of GM-CSF and FGF-2 [181]. In contrast, several studies found no association between high pregravid BMI and circulating TNFα [204, 230, 231, 233, 235, 236] IL-1β [175, 181, 225, 230, 237], or anti-inflammatory IL-10 [181, 223, 228, 229] during pregnancy.

Profiles of circulating chemokines have also been shown to vary with pregnancy [238, 239]. Significant shifts in chemokine profiles, notably MCP1 at the maternal-fetal interface, have been shown to trigger labor [240] and in some cases even initiate obstetric complications [239]. Pregravid obesity has been shown to elevate circulating innate immune cell recruiting factors CCL2 at various stages of pregnancy [181, 228, 231, 241] and CXCL8 at term [181]. There is emerging evidence of some cytokines such as IL-6 bidirectionally crossing placental barriers [242], however, whether similar patterns exist for other hormone/cytokines is still unclear.

Glucose tolerance and insulin resistance, often observed during pregnancy, are exacerbated with high pregravid BMI. Women with obesity exhibit significantly high levels of insulin at term, which vary linearly with BMI [181]. Similarly, other mediators associated with insulin resistance are also impacted by obesity. For example, leptin, a pro-inflammatory mediator that modulates appetite, implantation, and immune responses during pregnancy increases with gestation [243] and drops immediately after delivery [244, 245]. During pregnancy [246], serum leptin levels positively correlate with BMI and fat mass and remain significantly higher in obese mothers at term [181, 246, 247]. In contrast, while some studies report a positive association between pregravid BMI and circulating resistin levels [248, 249], others report no relationship [250, 251]. Taken together, these studies highlight the systemic impact of high BMI on chronic inflammation, immune activation, glucose tolerance, and insulin resistance.

Linking maternal obesity and fetal immune reprogramming – role of epigenetic mechanisms

The final section in this review summarizes mechanisms that link the maternal obesogenic environment with immune dysregulation observed in the offspring (Figure 4). Given the strong association between maternal pregravid obesity and evidence of systemic inflammation as well as in the liver, adipose tissue, and placenta, one can hypothesize that the fetus is exposed to a heightened inflammatory milieu. The placenta also harbors cells (immune and non-immune) that respond to conditions of oxidative stress and nutrient deficits by secreting various immune mediators, hormones, and growth factors into the fetal circulation. Newborns of mothers with obesity have relatively higher levels of leptin, and IL-6 in cord blood compared to newborns of lean mothers [224]. However, no studies to date have systematically measured maternal and cord blood cytokine levels in matched mother-baby pairs to assess potential transplacental transfer of inflammatory factors with the exception of IL-6 [242, 252].

Figure 4:

Mechanisms linking obesity associated changes on the maternal front and adverse immunological outcomes in the fetus and offspring.

Maternal lipids are also known to cross the placenta and contribute to fetal cholesterol synthesis, particularly during early gestation [253]. Fetal exposure to excess lipids, particularly saturated fatty acids, can activate pro-inflammatory pathways impacting nutrient metabolism, mitochondrial function, hematopoietic stem cell fate, and immune cell plasticity. Excess nutrients such as lipids can rewire cellular metabolism through epigenetic mechanisms, triggering transcriptional adaptations that ultimately alter immune cell function. Evidence of transcriptional reprogramming has been reported as early as the blastocyst stage in rats, with up-regulation of NF-κB regulated Ccl4 and Ccl5 in embryos of obese dams [254]. Pregravid maternal obesity has also been shown to alter expression of mitochondrial and lipid metabolism genes in human infant umbilical vein endothelial cells [255]. Stable changes in gene expression can be regulated via epigenetic mechanisms such as DNA methylation, and chromatin remodeling, and post-transcriptional regulation via microRNAs, which have been hypothesized as central to the fetal origins of adult health and disease [256].

DNA methylation patterns are established mainly during embryogenesis/early postnatal life. Within the immune system, DNA methylation play critical roles in hematopoiesis and can vary within cells of a particular lineage depending on their source (fetal liver, cord blood, bone marrow vs. peripheral blood) [257]. Additionally, methylation patterns overlapping enhancers and promoters of lineage-determining factors define differentiation trajectories of immune cells during development [257]. High maternal BMI is inversely correlated with methylation levels within the promoter of inflammatory genes in cord blood mononuclear cells [258]. Patterns of global loss in methylation seen in cord blood leukocytes from offspring born to mothers with obesity remained consistent up to three years of age [259]. Recent studies carried out using targeted bisulfite sequencing in purified cord blood monocytes demonstrated that pregravid maternal obesity is associated with DNA hypomethylation, particularly at promoters of transcription factors such as PPARG that are associated with a dampened transcriptional responses to LPS [43].

Maternal pregravid obesity is also associated with reduced abundance of the repressive histone mark H3K27me3 and increase expression of Zfp423, a key regulator committing cells to adipogenic lineage [260] within fetal adipose tissue in rodent models. Furthermore, altered expression of Wnt genes in livers of offspring born to dams on HFD show hyper-acetylation of Histone H4 and H3K9 [261]. Additionally, studies in Japanese macaques have demonstrated elevated acetylation at H3K9, H3K14, and H3K18 in fetal hepatic tissue exposed to HFD in utero [262]. Studies in animal models have suggested the potential role for histone deacetylases, mainly sirtuins in these epigenetic changes. Reduced SIRT1 activity in the fetal livers of macaque fetuses exposed to HFD during gestation were linked to altered expression of metabolic regulators PPARs and fatty acid synthases [263] as well as impaired liver mitochondrial SIRT3 expression [264]. Moreover, increased expression of the repressive methylation (H3K9Me3) was noted on TLR4 and LBP gene promoters in splenocytes of HFD-fed dams offspring contributing to poor responses to LPS stimulation [143]. These epigenetic “tags” could function as mechanisms of cellular memory within the fetal compartment in response to sustained maternal inflammation.

The precise series of events by which maternal environment/inflammation alter fetal epigenome are still unclear. However, it has been hypothesized that excess lipid exposure in utero [265] could impact cellular metabolism in immune precursor cells [162] thereby impacting epigenetic profiles of mature fetal immune cells. Additionally, inter-placental transfer of maternal inflammatory factors (including maternal cytokines and metabolic hormones) could establish immunological tolerance in fetal cells. Leptin, which functions as the nutritional signal in the fetus, is produced by the placenta and released into circulation. Some studies have reported elevated leptin levels in cord blood with higher maternal BMI [42, 224] where it is negatively associated with leptin promoter methylation [266]. Furthermore, a more recent study has demonstrated hypermethylation of LEPR (leptin receptor) promoter in the villi of mothers with higher BMI [267]. Whether these changes alter gene expression or surface expression of leptin receptor within cord blood immune cells would be an exciting avenue for future research, given the poignant role for leptin in both nutrient sensing and regulating immune responses.

Whether these epigenetic changes contribute to the adverse outcomes seen in offspring of obese mothers or if these cellular adaptations to sustained in utero exposure to inflammation is still up for debate. Future work in animal models will have to test if changes in fetal cells can be linked to specific changes in offspring immune function. Nevertheless, these studies demonstrate the importance of epigenetics in intergenerational health and provide a potential mechanism by which maternal obesity could program fetal immune function.

CONCLUDING REMARKS

Despite the recent progress made assessing the specific burden of maternal obesity on several compartments of fetal immunity, the specific sequences of events that link maternal inflammation with fetal outcomes is still unclear. Several other questions still remain: Is there a specific window during gestation when the fetus is most susceptible to maternal obesogenic environment? Are there cell extrinsic factors (cytokines, hormones, defective cell to cell communication) that contribute to dysregulated immune responses in the fetus? Are there factors that cross the placental barrier skewing fetal immune cells to a refractory state or is it just an outcome of poor immune development? Are these adverse outcomes in the offspring reversible with maternal pre-pregnancy weight loss interventions or therapies? While large human cohort-based studies are able to identify specific defects in immune responses, the genomic and epigenomic determinants and more importantly causality can only be tested in well-designed animal models. These models can to a large extent overcome the complexity of homeostatic gestational immune adaptations both in the blood and in the placenta, and genetic and environmental confounders in human subjects. Deep profiling of immune cells using a combination of single cell technologies will expedite and provide a more accurate assessment of the immunological outcomes to specific perturbations. Elucidating these molecular underpinnings will be crucial in designing new effective therapeutic strategies or leveraging existing therapies for amelioration of maternal obesity-associated changes in offspring immunity.

Abbreviations

BMI

Body Mass Index

WD

Western Diet

HFD

High Fat Diet

PE

Preeclampsia

GH

Gestational Hypertension

GD

Gestational Diabetes

NHP

Non Human Primate

HSC

Hematopoietic Stem Cel

(p/m) DC

(plasmacytoid/myeloid) Dendritic Cell

NK

Natural Killer

MODM

Monocyte-Derived Macrophage

TLR

Toll-like Receptor

NLR

NOD-like Receptor

APP

Antimicrobial Proteins and Peptides

PRR

Pattern Recognition Receptor

MDSC

Myeloid-Derived Suppressor Cell

APC

Antigen Presenting Cell

ROS

Reactive Oxygen Species

CRP

C-reactive protein

LPS

Lipopolysaccharide

HLA

Human Leukocyte Antigen